300 Rev. P. J. Kirkby : Clwmkal Effects of the 



positive column " is not sufficiently intense to cause the 

 negative ions that pass through it to ionize the gas except to 

 a small extent. 



Thus in the case of a gas where no chemical action takes 

 place far fewer positive ions than negative pass through the 

 positive column per second. Hence any chemical action that 

 is observed in the positive column when a discharge passes 

 through hydrogen and oxygen should on our hypothesis be 

 attributed to the motion of negative ions, unless the chemical 

 union of these gases is attended by the generation of ions. Now 

 it is true that the explosive union into water of hydrogen and 

 oxygen is attended by the formation of ions; but the number of 

 ions formed is incomparably smaller than the number of mole- 

 cules of water (C. E. Haselfoot and P. J. Kirkby, Phil. Mag. 

 Oct. 1904). We conclude then that any chemical action 

 that occurs in the positive column is exclusively due to the 

 impacts against the gaseous molecules of the negative ions. 

 But before estimating the chemical action per centimetre of 

 the positive column, it is convenient to investigate the force 

 there upon which depends the motion of the ions. 



Estimation of Electric Force in the Positive Column. 



To determine approximately the electric force in the 

 positive column during experiments made at the same mean 

 pressure, curves have been drawn having as abscissas the 

 distances between the electrodes and as ordinates the mean 

 potential difference in volts of the electrodes. Each of these 

 curves applies to one pressure and was constructed from the 

 numbers in one of the tables. But to avoid confusion only 

 four of them are shown in fig. 3, and these four are each 

 derived from the same table which contributed one of the 



each kind within it are always almost exactly equal. For the constancy 

 of the force in the positive column requires that the stream of negative 

 ions has the same density as the stream of positive ions. The former 

 probably moves with a much greater velocity than the latter, since the 

 pressure is so low ; hence the number of negative ions which traverse 

 the positive column per second is, in the same proportion, greater than 

 the number of positive. Thus if u, v are the velocities of the streams 

 of negative and positive ions, C 5 , C 2 the currents in amperes carried by 

 these streams, so that C^+C,, is the whole current, and if A is the cross- 



section of each stream, the equation t~ =4rrp becomes in the positive 



column -?— x I — - 1, X being m volts per cm. Hence 



if X is nearly constant, C l : C 2 = u : v almost exactly. I assume in what 

 follows that no positive ions are generated near the anode, the expla- 

 nation does not essentially depend on that assumption at all (see below) . 



